CUDA error: operation not supported

[Pancat009]

CUDA error: operation not supported
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

my env:
CUDA:12.4
Python version: 3.10.11
pytorch version: 2.4.0+cu124
xformers version: 0.0.27.post2

[Pancat009]

CUDA: 12.6 reports the same error

[Pancat009]

if you are using 秋叶版aki-ui, try to switch the CUDA memory allocation
change it to native